INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master- UMI films the text directly from the original or copy submitted. Thus some thesis and dissertation cupies are in typewriter face, while others may be from any type of computer printer.
The quality al this I.produc(kn k dependent upon th. quality of the copy submW. Brdcen or bldistinct priM cokred or poor quality iltustralions and photographs, print Meedthrough, substandard margins, and improper alignment can adversely afkt reproduction.
In the unlikely event that the author did not send UMI a cornpkte maiptand there are missing pages, these will be noted. Also. if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charto) are reproduced by S8dl*oning the original, beginning at the upper left-hand comer and continuing from left b right in equal sections with small overlaps.
Photographs included in the original manudpt hwe been mpoducad xerographically in this copy. Higher quality C x 9" black and white photographic prints are available for any photographs or illustratio~appearing in this cwfw an additional charge. Contact UMI directly to der.
Bell & Howell Information and Learning 300 Norlh Zeeb Road, Ann Arbor, MI 48106-1346 USA
NOTE TO USERS
This reproduction is the best 'copy available
National Library BibliitMque nationale 1+1 of,,, du Canada Acquisitions and Acquisitions et Bibliographic Services services bibliographiques 39s weuiiqtm street 395, rue Wellington Ottawa ON KIA ON4 OttawaON KIA ON4 Canada Canada
The author has granted a non- L'auteur a accorde une licence non exclusive licence allowing the exclusive pernettant a la National Library of Canada to Bibliotheque nationale du Canada de reproduce, loan, distriiute or sell reproduke, preter, distrr-buer ou copies of this thesis in microform, vendre des copies de cette these sous paper or electronic formats. la forme de microfiche/fb, de reproduction sur papier ou sur format electronique.
The author retains ownership of the L7auteurconserve la propriete du copyright in this thesis. Neither the droit d'auteur qui protege cette these. thesis nor substantial extracts &om it Ni la these ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent &e imprimes reproduced without the author7s ou autrement reproduits sans son permission. autorisation. INTEL AND INTRASPECIFIC PHYLOGEOGRAPHY OF NORTH AMERICAN ODOCOLLEUS DEER BASED ON MITOCHOND~DNA SEQUENCES
by
O Annette D. Greenslade, B.Sc. @onours)
A thesis submitted to the School of Graduate Studies in partial llfillment of the requirements for the degree of Master of Science
Department of Biology Memorial University of Newfoundand
July 1998
St. John's Newfoundland ABSTRACT
Phylogenetic analysis ofmitochondrid DNA (mtDNA) cytochrome b sequences identified four major mtDNA genotype assemblages among popuIations of Odocoileus in western North America. These assemblages correspond to northwestern black-tailed deer, southwestern mule deer, northwestern mule deer, and southwestern white-tailed deer. Approximate times of the divergence of these assemblages, calculated on the basis of pairwise nucleotide &quence divergence estimates, were used to construct a model of
Odocoileus evolution in western North America. According to this model, black-tailed deer represent the ancestral OdocoiZ~mtDNA lineage. Mule deer and white-tailed deer diverged more recently in palaeontologicaI history. The southwestern white-tailed deer mtDNA genotype assemblage is more closely related to the mule deer mtDNA assemblages than it is to the southeastern white-tailed deer Lineage. This suggests that the southwestem white-tailed deer mtDNA lineage may have been effectively replaced by mule deer mtDNA through relatively recent hybridbation between the two species.
Microgeographic analysis of deer fiom southern Alberta revealed a considerable degree of geographic structuring of mtDNA sequence genotypes, as did qualitative analysis of the geographic distribution of mule deer and white-tailed deer mtDNA genotypes in
CalSornia and Alberta. The phylogeographic structure may be maintained by philopatry of the female (or family) social units of Odocoileus deer, despite their potentially high vagility . Gratitude is extended as appropriate to my supervisory committee (S. M. Carr,
D. I. Innes, and E. H. Miller), my colleagues, and those researchers who provided deer samples: S. Lingle collected samples fkom the McIntyre Ranch near Mapth,
Alberta, and the Waterton Lakes National Park, Alberta; M- Cronin (LGL Alaska) provided samples from southern California; C. Strobeck (University of Alberta,
Edmonton) provided samples fiom British Cohunbia, Manitoba, Sasketchewan, and
Ontario; and S. Carr &Iemorial University of Newfoundlaud, St. John's) provided samples from western North America. DNA extractions of the samples from S. Lingle were carried out by J. Coffin et 01. of C. Strobeckrs research lab. S. Lingle also provided a UTM map showing the sampling location of most of the deer she collected;
Figure 2 is based on this map.
Many thanks to M. Dennison and A. Lynch (both at The Royal Ontario Museum) for help with preparing the maps, to 0. Haddrath (The Royal Ontario Museum) for advice with phylogenetic analyses, and to A. J. Baker me Royal Ontario Museum) for reading an earlier draft of my thesis and offering invaluable advice (thanks for the encouragement).
Thank you to my supervisory committee for reading drafts of the thesis and for offering constructive criticism. Thanks also to my supervisor for help with using
MEGA, REAP, and PAUP,and with developing a model of 0docoUeu.s evoIution.
ALSO sincere gratitude is extended to G. Kemy, Graduate secretary of the Biology
Department, for help with all the administrative details. TABLE OF CONTENTS
Page - Abstract ...... n-
Acknowledgements ...... rn-.-
Table of Contents...... v . List of Tables.-...... ~...... vu-
List of Figures ...... wu--.
Introduction...... 1
Materials and Methods...... 13
Results.,.,...... -...... 23
Microgeographic ana&sis ofdeer#om Magruth and Waterton Lakes National Park...... 23
New rntDNA sequence genowes identtperi:...... 3 0
Comparison ofpreviously reported mtDNA sequence genotpes ond new &ta ...... 32
Pattern of mcleotide substitution among Odocoileus mtDNA sequence genows..~..~~~...... ~...... 33
Phylogenetic analyses of mtDNA sequence genotypes ...... 44
Discussion.-..,-...... ,....-....-..-...... 78
Palaeontological record of Odocoileus...... -...... --78 Evohtionory history of Odocoileus bared on morphological
and mtDNA sequence &ta ...... RR-R-RRRRRRRR--RRRRRRR~ ...... 79
Glacial refirgia, stochastic lineage sorting d introgressive hybridization of mtDNAA-...... 86
Gene trees and species trees...... 91
Phylogeogap?iy of Odocoileus~..~~...... 93
References...... 97 LIST OF TABLES
Table
1 Mitochondria1 DNA genotypes found in deer fiom the Mctntyre Ranch near Magrath, Alberta, and Waterton Lakes National Patk, Albem...... -...... -.... d
2 Pairwise haplotype divergence estimates for the mtDNA genotypes identified in mule deer and white-tailed deer fkm the McIntyx Ranch near Magrath, Alberta, and Waterton Lakes National Park, Alberta...... 28
3 Indices of haplotype (h) and nucleotide (x) diversity within populations of mule deer and white-tailed deer fiom the McIntyre Ranch near Magrath, Alberta, and Waterton Lakes National Park, Alberta...... 29
4 Correspondence of 401-nucleotide sequence genotype with 307-nucleotide genotype (Carr and Hughes, 1993; Hughes, 1990; Hughes and Cam, 1990), and with RFLP- genotype (Cronin, 1991; Cronin and Bleich, 1995)...... 3 1
5 Frequencies and distributions of variant nucleotides in 0docoileu.s within the 40 1-nucleotide cytochrome b sequence data...... 43
6 Minimum and maximum pairwise nucleotide sequence divergences within and among OdocoiIeus mtDNA genotype assemblages...... ,...... 70
vii LIST OF FIGURES
Figure
1 Geographic distribution of Odocoileus virginiantrs (white-tailed deer) and 0. hemiom (mule deer and black-tailed deer) in Noah America ...... 3
Sampling Location of white-tailed and mule deer mtDNA genotypes within the southern Alberta collection site (Mchtyre Ranch near Magrath, and Waterton Lakes National Park)...... C.C...C....26
Mitochondrial cytochrome b sequences of genotypes reported in Odocoileus...... 35
Character matrix of the variable nucleotide sites used in the phylogenetic analysis of Odocoilellr relationships...... -41
Neighbor-joining tree constructed fiom Kimura-2 parameter pairwise distance estimates (transitions + transveersions)...... -46
Neighbor-joining tree constructed using synonymous substitutions with a Jukes and Cantor correction...... 48
Neighbor-joining tree constructed using nonsynonymous substitutions with a Jukes and Cantor correction.*...... ,., ...... *.ti*..50
Quartet puzzling tree of the phylogenetic relationships among Odocoileus mtDNA sequence genotypes...... -52
Maximum parsimony 50% majority rule consensus tree of the 106 minimum length trees recovered by a heuristic search of the unmodified mtDNA sequence genotype data...... -55
viii Unrooted minimum-length network of phylogenetic relationships among mule deer mitochondrial DNA genotypes in North America ...... 60
Unrooted minimum-length network of phylogenetic relationships among white-tailed deer mitochondrial DNA genotypes inNorth America...... 62
Unrooted minimum-length network of phylogenetic relationships among mule deer and white-tailed deer mitochondrial DNA genotypes in North America ...... 64
One of two umooted minimum-length networks of phylogenetic relationships among mule deer, white- tailed deer, and black-tailed deer mitochondrial DNA genotypes in North America...... 66
The second of two unrooted minimum-length networks of phylogenetic relationships among mule deer, white- tailed deer, and black-tailed deer mitochondrial DNA genotypes in North America...... 68
Approximate distribution of Odocoileus hemionus mtDNA genotypes h California...... 72
Approximate distribution of Orlomileus hemionus mtDNA genotypes in Alberrta.,...... 74
Approximate distribution of Odocoilercs virginianus mtDNA genotypes in Alberta...... 76
Model of evolutionary pattern among Odocoileus mtDNA assemblages in North America...... 85
Extent of ice coverage over northern North America during the Wisconsin Glaciation...... -...... 90 The genus Odocoiieus comprises two extant species, 0-virginianus (whitetailed deer) and 0. hemiorua (mule deer and black-tailed deer). Grubb (1993) places them within the subfamily Capreolinae (family Cervidae, order Artiodactyla), whereas most authorities place them within their own subfamily, the Odocoileinae (Whitehead, 1972;
Baker, 1984; Miyamoto et al., 1993). The two species are sympatric over much of their range west of the Rocky Mountains (Figure 1); the 30 subspecies of 0. virginianw currently recognized maker. 1984) are broadly dism'buted in mesic and tropical environments throughout North America and extend into Centnl and South
America (Mccullough, 1987), while the seven 0. hemi0tw.s subspecies (WaIlmo, 1981) occupy a more restricted range in the xeric environments of western North America, from southeastern Alaska to western Mexico. The distribution of 0-h- cotuMiantcs
(Colurnbian black-tailed deer) extends from central California to southern British
Columbia, and is adjacent to the distribution of 0. h. sitkensis (Sibblack-tailed deer), which extends northward along the coast of British Columbia to southeastern Alaska
(Wallmo, 1981). AU other subspecies of 0. hemionus are referred to as mule deer, and occupy the remainder of the geographic area identiki in Figure 1 as the 0. hemionus species range. Subspecies of 0. hemionus are characterized by Figure I: Geographic distribution of Odocoilcus virginhm (white-tailed deer) and 0-hemionus (mule deer and black-tailed deer) in North America- The map was assembled by S. Cam, based on Baker (1984). Cross- hatching indicates the area of sympatry between the species. discontinuities in characters that result fkom barriers to gene flow, such as deserrs and mountain ranges. In contrast, geographic variation in white-tailed deer is clinal, and differences among subspecies seem to be largely arbitrary (McCulIough, 1987). The two species, however, can be readily distinguished on the basis of several characteristics, including those of morphology, ecology, and behaviour (Baker, 1984;
Marchinton and Hirth, 1984). Some of the morphological featmes that differentiate the two species of Odocoileus are antler shape, length and position of the metatarsal gland, and appearance of tail, ear, and rump patch. For exampIe, the antler tks bifurcate in mule deer, but not in white-tailed deer, and the beams tend to curve forward and turn inwards more in the latter species. Wbite-tailed deer have longer tails of Lighter colour. Metatarsal gland characteristics are considered to be the most reliable in distinguishing the two species (Derr er al., 1991). The metatarsal ridge of white-tailed deer is considerably shorter (25 mm) relative to that of mule deer (125 mm), with no overlap in range; the metatarsal gland of white-tailed deer is located below the midpoint of the shank, whereas it is above the midpoint of the shank in most mule deer; and the metatarsal tuft is white in co1our in white-taiIed deer, and brown in mule deer (Wishart, 1980).
White-tailed deer and mule deer are regarded as reproductiveIy isolated species
(WalImo , 198 I), although evidence suggests that hybridization and introgression of genetic material occurs between species or subspecies (Ballinger et al., 1992; Cam sad
Hughes, 1993; Cam et al.. 1986; Hughes and Cam, 1993; McCIymont et aL, 1982;
Wishart, 1980). Deer with intermediate morphological characteristics have been produced in captivity, and reported in the wild in western North America. Wishart
(1980) indicated that hybrids in Alberta are intermediate in some characters and more like one of the parental stocks in others. The tails of all hybrids are dark, shading into black towards the tip. The intermediate location and size of the metatarsal gland and the conspicuous white or mostly white hairs surrounding the gIand are typical hybrid characters.
The occurrence of hybrid individuals has prompted a number of studies of
Odocoileus, particularly at the genetic and behaviowal levels. White-tailed and mule deer differ in habitat preference, with white-tailed deer preferring wetter, more wooded areas that offer much cover, whereas mule deer prefer more open and rugged areas
(Kramer, 1973; Swenson et ai.. 1983; Smith, 1987). This difference does not appear to be the result of competition for food resources, despite similar food preferences, nor does it seem to be caused by direct behavioural interactions. Geist (1981) suggested that the different antipredator responses of white-tailed and mule deer underlie their habitat segregation. White-tailed deer avoid predators at large distances and gallop when threatened, whereas mule deer approach disturbances and avoid predators by means of an escape gait described as a stott. The escape gait of captive hybrid - individuals, described as a bound, is intermediate between that of each of the parr&l species; it is slow and mechanically inefficient, which may result in ineffective responses to predators in the deer's natural habitat. This could in turn lower the chances of hybrid Survival relative to white-taiIed and mule deer, thus restricting contemporary genetic interchange between the two species (Lingle. 1992; LingIe
1993).
Many genetic studies of Odocoilew have investigated interspecific and intraspecific genetic variation. Several of these studies have involved analysis of the rmclear genome using protein electrophoresis. One of the earliest such seeswas an analysis of blood serum protein variation between 0. virginionus and 0. hemiom, aud among subspecies of 0.hemionw (Cowan and Johnston, 1962). Othrr allozyme studies have focussed on Odocoileus populations in specific geographic locations, such as South
Carolina (Cothran et al., 1983; Ramsey e? af-, 1979). Tennessee ennedy ef al. ,
L987), Maryland (Sheffield et 41.. 1985). Texas (Stubblefield et al. , l986), Alberta
(McClymont et al., 1982). Oregon and Washington (Gavin and May, 1988). and the southwestern United States @err, 1991).
A second approach has been the analysis of the mitochondria1 genome, either as the primary source of genetic information or in conjunction with allozyme analysis. Mitochondria1 DNA ONA)has been used in a kge number of studies to investigate aspects of genetic diversity, phylogeography, population structure and dynamics, and population evolution. Animal mtDNA is a double-stranded, covalently closed circular molecule found within the mitochondria. It has a conserved gene content, consisting of two ribosomal RNA genes, 22 transfer RNA genes, and 13 protein genes which encode enzyme subunits that function in electron transport or ATP synthesis (Wilson ef oC., 1985; Moritz et ul., 1987). Animal mtDNA lacks introns, repetitive DNA, and pseudogenes; intergenic sequences are small or absent (Avise et al., 1987; Moritz et al., 1987).
The popularity of mtDNA in evolutionary studies is due to a number of characteristics of the molecufe, induding its small size, high copy number, the relative ease with which it can be isolated from nuclear DNA, its fast rate of sequence evolution, conserved gene content, maternal inheritance, effeaive haploidy, and lack of recombination (PaIumbi and Baker, 1994; Slade et al., 1993; Zhang and Hewitt, 1996).
The availability of universal primers for use in mtDNA amplification via the polymerase chain reaction (PCR), and the existence of a large data set for use in comparative studies have also made mtDNA the molecule of choice for many studies
(Palumbi and Baker, 1994; Moritz et ai., 1987).
Although the miparental inheritance of mtDNA makes it an effective marker in studies of hybridization and maailimal gene flow (Carr and Hughes, 1993; Hughes and
Cam, 1993; Avise. 1994). it does not allow the reconstruction of paternal lineages (Wilson et al., 1985; Avise et af- , 1987). This limitation may not affect studies of species that do not exhibit gender-biased dispersal, but will be important if the species under study have compIex social behaviouzs and differences in the dispersal of males and females. In such cases, the population structure as indicated by matemally inherited mtDNA may be qyite differeat fkom that indicated by biparentally- inherited nuclear DNA (Palumbi and Baker, 1994). However, comparisons of the spatial distribution of nuclear and mitochondria1 markers can be used to study sexual bias in dispersal (Moritz et al., 1987).
Because mtDNA does not undergo recombination, the evolution and inheritance of mtDNA are less complicated to decipher, relative to nuclear DNA. However, the entire mitochondrial genome must be regarded as a single genetic locus (SIade et al.,
1993; Palumbi and Baker, 1994); in the absence of recombination, the 37 genes effectively behave as a single genetic entity (Wilson et al., 1985) that is passed on to progeny in its entixety (Attardi, 1989.
Other potential complications in studies based on mtDNA are the detection of nuclear copies (pseudogenes) of mitochondrial sequences, and paternal leakage of mtDNA. If mitochondriallyderived sequences are present in the nuclear genome, and a total genomic DNA extract is used as the initial DNA source in PCR, the nuclear copies may co-amplify, thereby causing ambiguities in the data that appear as sequence heteroplasmy and genetic polymorphism (Zhang and Hewitt, 1996). Gyllensten et al.
(1991) showed that a low proportion of paternal mtDNA could be produced in heteroplasmic experimental hybrids by repeated backcrossing between mouse species
Mus rmcsculdls and M' domesticas. If the OCCWfence aad degree of paternal leakage of mtDNA were great enough, mtDNA molecules could potentialiy recombine, thereby complicating the reconstruction and interpretation of mtDNA-based genealogies (Avise,
199 1). Paternal leakage of mtDNA does not seem to be a factor in most studies, however.
Genetic studies of Odocoileus have reported low levels of hybridization in deer populations born Montana (Cronin et al., 1988). Alberta (Hughes, 1990; Hughes and
Cam, 1993 ;McClymont et ai., 1982), and the Pacific Northwest (Cronin, 199 1; Gavin and May, 1988), and a high level of hybridization in Texas (Ballinger et al.. 1992;
Carr et al., 1986; Stubblefield et al., 1986). With regards to mtDNA genotypes identified by direct sequence analysis, there are several cases of a genotype typical of one species being found in individuals that are phenotypically the other species.
Hughes and Carr (1993) report one mule deer individual from Alberta that had a white- railed deer mtDNA genotype (EDN), and one white-tailed deer, also &om Alberta, that
9 had a mule deer mtDNA genotype (ELP). Such evidence suggests reciprocal hybridization between the two species.
The term "phylogeography" was tifit used by Avise et ai. (1987) aad is discussed in greater detail by Avise (1994). Avise (1994) has defined phylogeography as "the study of the principles and professes governing the geographic distributions of genealogical lineages, including those at the intraspecific leveln @. 233). The application of modern molffalar techniques, especially those that use mtDNA, to investigations of population genetics has prompted the development of the discipline of intraspecific phylogeography. The advantageous characteristics of mtDNA (maternal, non-recoinbining mode of inheritance, and rapid rate of sequence evolution) often produce multiple alleles or genotypes that can be ordered phylogenetically within a species, thus producing an ineaspecific phylogeny or gene genealogy. Studies have indicated that rntDNA clones and clades of some species are geographically structured, thus illuminating the link between phylogeny and geography, and therefore between evolution and the physical environment.
The study of mtDNA as a molecular marker allows maternal lineages to be traced through space and the, and thus pennits reconstruction of the evolutionary history of a species. The present-day distri'bution of any given species represents the product of many factors, including those inherent in the environment and in the organism itself- Physiogeographic factors such as mountain ranges and river systems may act as barriers to dispersal, while the vagility of the organism will also tie of importance in establishing the geographic dism'buton of a species. It is expected that highly vagile organisms will not show a great deal of phylogeographic structure. Although this is the case for animals such as black bears (Ursus mnen'cantcs) (Cronin et a[., 1991) and coyotes (Cm3 larrm) ahmanand Wayne, 1991; Lebman a uL, 1991), populations of other species, such as humpback whales (Megoptera mvaeangliae) (Baker et al.,
1990) are subdivided, as indicated by localized genealogical structure and/or major gaps in the rntDNA phylogeny across the species range (Avise, 1994).
Behavioural characteristics, such as philopatry and dispersal, will also affect the level of phylogeographic structure exhibited by a given species. While the social organization of a species may restrict dispersal, particularly if it involves gender-biased dispersal, its effects may not be obvious or detected in molecular studies of the nuclear genome. Such behaviour would appear to be of greater consequence when the molecular marker used is found within the mitochondrial genome; if the females of the species exhibit philopatric bebaviour with respect to breeding, and the social organization involves the establishment of female-based family groups, as in
Odocoileus, the mtDNA phylogeny is expected to indicate a high degree of spatial- genetic structure. The current study analyzes the phylogenetic relationships among Odocoifeus mtDNA genotypes in western North A.mer&, and interprets these reiatiomhips within a biogeographical context. Whereas previous genetic studies of Odocoiieus have focused on specific populations or geographic regions within North America, the current study attempts to combine and extend existing mtDNA data sets and consider the large-scale inter- and intraSpeeific phylogenetic relationships within Odocoileus. SAMPLE COLLECTION
Samples used in this study came from a variety of sources. A set of white-tailed deer and mule deer tissue (ear notches or ear punches) andlor blood samptes &om fawns on the McIntyre Ranch near Magrath, Alberta, -andWaterton Lakes National Park,
Alberta, was collected (S. Lingle). Additional samples were fmm deer that were found dead, either as a result of predation or road accidents. Samples were colIected between
October 1993 and October 1994, with most firom July 19W. The extraction of DNA from these samples was performed by C. Strobeck's lab, University of Alberta,
Edmonton, Alberta, Canada. A sub-set of duplicate &sue samples were also sent directly to me.
Additional samples representing other geographic locations were also obtained.
These included a small set of DNA extracts from deer collected in British Columbia,
Manitoba, Saskatchewan, and Ontario (C. Strobeck); a set of mule deer DNA samples from southern California consisting of DNA from two representative individuals of each of eight mitochondrid DNA genotypes as identified by restriction-fragment-length- polymorphism (RFLP) analysis (Cronin and Bleich, 1995); and representative DNA samples for each of a series of previously identified cytochrome b mtDNA genotypes within white-tailed, black-tailed and mule deer in western North America (Cam and
Hughes, 1993; Hughes and Carr, 1993). A single moose (Alces alces) sample was also
provided by S. Cam for use as an outgroup in phylogenetic analyses.
DNA EXTRACTION
The majority of samples were provided in the form of extracted DNA. Samples
provided by C. Strobeck were extracted with a commercial kit, the QIAamp Tissue Kit
(QIAGEN Inc., Chatsworth, CA), according to manufacturer's iastructions. DNA
samples fiom M. Cronin and S. Carr were extracted by standard techniques as described
in Cronin and Bleich (1995) and Carr and Marshall (1 99 1).
For those samples provided in tissue form, DNA was extracted using the acid quanidinium thiocyanate-phenol-chloroformprocedure of Chomczynski and Sacchi
(1987) as modified by Bartlett and Davidson (1 99 1). Approximately 100-200 mg of tissue were homogenized using a sterile plastic pestle in 450 pL of a solution containing
4 M guanidinium thiocyanate, 25 mM sodium citrate (pH 7.0), 0.5% SarkosylR,and 0.1
M 2-mercaptoethanol. A 50 pL volume of sodium acetate (2 M, pH 4.1) was added to the homogeaate, followed by 300 pL of Tris-saturated phenol, and 150 pL of
chloroform/isoamyl alcohol (24: 1, vfv). The solution was mixed and left on ice for 15
minutes. Following this incubation, the sample was centrifbged at 10 000 x g for 15
minutes at 4OC, after which the upper (aqueous) phase was transferred to a new 1.5 mL microfigetube, 450 pL ofchloroform~isoamylaIcohol(24:1, vfv) was added, and the solutions were &ed gently by inverting the tube. The sample was centrifbged at 10 000 x g for 15 minutes at 4OC, after which the upper (aqueous) phase was transferred to a new
1.5 mL microfuge tube and 400 pL ofcold isopropanol was added. The solutions were mixed by inverting the tube and were left at -20°C for at least two hours (or overnight) to precipitate the nucleic acids. The sample was then centrifbged at 10 000 x g for 15 minutes at 4OC, and the isopmpanoI was removed and discarded, leaving the nucleic acid pellet in the tube. A 150 pL volume of kshice-cold (-20°C) 75% ethanol was added to the tube, and was followed by another centrifugation step (10 000 x g, 15 minutes, 4 OC).
The ethanol was removed, the pellet was dried briefly (approximately 10 minutes) under vacuum and was resuspended in 100 pL of sterile distilled water.
AMPLIFICATION OF DNA
Amplification of a segment of the mitochondrial DNA genome was carried out using the polymerase chain reaction (PCR). The sequences of the primers used to amplify and sequence a 401-nucleotide fragment of the cytochrome b gene are as follows:
S-GCCCCTCAGAATGATA'ITTGTCCTCA-3'(HI 5 149) (modified fiom Kocher et al.,
1989) and 5'-CGAAGCnGATATGAAAAACCATCGTTG-3'(L 14724) (Inrrin ef al., 199 1). Primers were synthesized by the Oligonucleotide Synthesis Laboratory, Queen's
University, Kingston, Ontario, Canada
Amplifications were carried out in 100 pL volume reactions containing 67 mM
Tris-HC1 (pH 9.0 at 2S0C), 2 mM MgCl, 10 mM 2-mercaptoethanol (all Sigma), 0.2 mM each of dATP, dCTP, dGTP, and drrP (Pharmacia), 10 pmol each of the heavy- and light-strand primers, and I unit of Amplie DNA poIymerase (Perkin-Elmer Cetus).
For each reaction, 2 pL of DNA extract were added. Most samples were amplified in a
Perkin-Elmer Cetus GeneAmp 9600 Thermal Cycler with the following step-cycle profile: an initial 5 minute denaturation at 95OC, followed by 35 cycles of 93OC for 15 seconds, 40°C for 30 seconds, 55°C for 30 seconds, and 72°C for 45 seconds. This was followed by a final extension step of 72" for 2 minutes. The standard PCR cycle consists of a denaturation step, followed by a single annealing step, and then an elongation step.
The cycle used here uses two annealing steps (40°C for 30 seconds and 55°C for 30 seconds) in order to prevent disassociation of the primer hmthe template DNA between the annealing and elongation phases. The intermediate step between the annealing and elongation steps makes the increase in temperature more gradual. An alternative approach is to use a temperature ramp. Some samples were amplified in a Perkin-Elmer
Cetus TC-1 DNA Thermal Cycler using an equivalent cycle: an initial 5 minute denaturation at 9YC, followed by 35 cycles of 93°C for 1 minute, 40°C for 1 minute, 55°C for 30 seconds, and 72OC for 2 minutes. The final extension step was 72°C for LO minutes. The differences in cycle profile reflect differences in the respective thermal cyclers; PCR products obtained using each machine were of comparable quality. The reaction mixtures were the same for both thermal cyclers, except that a drop of light mineral oil (Sigma) was added to each sample ampIified in the TC-1 Thermal Cycler in order to prevent evaporation ofthe sample during amplitication.
Following amplification, a 5 pL podon of the product was added to 1 pL of 5x tracking dye (25% glycerol [I -26g/d s~oc~]~ 50 mM N%EDTAy0.5% SDS, 0.1% bromophenol blue) and subjected to electrophoresis through a 2% NuSiev6 GTP agarose (FMC) gel in lx TBE buffer (pH 7.4). The presence of amplified DNA was confirmed by visualization of the DNA using ethidiurn bromide (1 pg/mL) under 302 nrn
UV illumination. A molecular weight standard (Haem digest of @ X phage DNA;
Pharmacia) was also run on the gel to estimate the size of the product and ensure amplification of the appropriate fhgment. Direct purification of the product was carried out using the Magi* PCR Preps DNA Purification System (Promega Corp.) according to manufacturer%instructions. The purified DNA product was quantified using a Hoefer model TKO 100 DNA Fluorometer (Hoefer Scientific Instnunents) and a 250 pg/mL calf thymus DNA (Clontech) weight standard reference. DNA SEQUENCING
The DNA sequence of each sample was determined using an plied Biosystems model 373A automated DNA sequencer and the Applied Biosystems PRISMRReady
Reaction DyeDeoxy Terminator Cycle Sequencing Kit. Reactions were set up according to manufacturer's instructions. A total of400 ng of template DNA were used, which corresponds to 3.2 pmol of DNA. The amount of template DNA required in the sequencing reaction depends on the size of the fragment; equimolar amounts of primer and template DNA produce optimal results. The DNA was completely dried under vacuum prior to the addition of the donmixture! (9.5 pL of the manufacturer's terminator premix, 3.2 pL (3.2 pmol) of primer, and 7.3 p.L of steriIe distilled water).
The sequencing primers were the same primers used to amplify the target DNA (I315149 and L 14724). Cycle sequencing of all samples was performed in a Perkin-Elmer Cetus
TC-1 Thermal Cycler using the following stepcycle profile: 98OC for 1 second, 50°C for
15 seconds, and 60°C for 4 minutes, for a total of 2S cycles.
The samples were then purified by means of SephadexRGSO Fine (Pbarmacia) spin column purification. Columns were prepared by Wgempty columns (5 Prime -> 3
Prime) with SephadexR,placing the columns in 1.5 mL collecting tubes, and centrihging
for one minute at 1,100 x g in a clinical centrifuge. The excess water was removed from the collecting tubes, the columns were placed in the same set of tubes and centtifuged for 30 seconds at 1,100 x g to remove any remaining water- The water was removed fiom the collecting tubes and the entire cycle sequencing product was carefdIy loaded onto the center of the SephadexRresin bed (one sample per column). Each column was placed in a sterile labelled 1-5 mL collecting tube, and was then cenmged for 3 minutes at 1,100 x g to collect the DNA. The DNA samples were completely dried under vacuum, resuspended in 4 pL ofa mbaure of 5pL of formamide and t pL of 50 mM Na2EDTA, heated to 9S°C for 2.5 minutes, and then chilled immediately on ice prior to being loaded on a 6% polyacrylamide denaturing gel. Gels were run for 8 hours at 32W on the 373A automated sequencer (Applied Biosystems).
DATA ANALYSIS
Data generated by the automated sequencer were collected and analyzed using the Data
Collection (v. 1.2), Analysis (v. 1.2) and SeqEdTM(v. 1.2) (Applied Biosystems, Inc.) programs. The DNA sequence data were exported to and aligned using the Eyeball
Sequence Editor (ESEE) (version 2.00) program of Cabot and Beckenbach (1989).
The Molecular Evolutionary Genetics Analysis (MEGA) program (v. 1.O 1) (Kumar et 01..
1993) was used to compute painvise distance estimates (transitions + tnursversions) under the Kimura 2-pantmeter substitution model (1980), and a neighbor-joining tree (Saitou and Nei, 1987) was constructed using the distance matrix. The moose, Akes alces, was used as an outgroup. Neighbor-joining trees were also constructed in MEGA using only synonymous substitutions, and using only nonsynonymous substitutions, both with a
Jukes and Cantor (1 969) correction.
Puzzle, a maximum likelihood-based quartet puzzling program (Strimmer and von
Haeseler, 1996), was also used in phylogenetic reconstruction. The type of analysis selected was tree reconstruction using the quartet pegtree search procedure over
1000 puzzling steps. Akes aices was used as an outgroup. The Tarnura and Nei (1993) model of nucleotide substitution was selected, Puzzle assessed the robustness of the reconstructed tree via reliability percentages (i.e., the number of times a group is reconstructed during the puzzling steps).
Cladistic analysis was performed on the sequence data using the Phylogenetic Analysis
Using Parsimony (PAW) (v.4.OdS 1 and subsequent versions) program (Swofford, 1996).
The heuristic search algorithm was used to identify maximum parsimony trees, and analysis was carried out using ten random stepwise additions of taxa and the nearest- neighbour interchange branch-swapping option. Branches were collapsed, creating polytomies, in any cases where the maximum branch length equaled zero. AU nucleotide substitutions were weighted equally, and Alces ulces was used as the outgroup in the initial heuristic search that was performed on d taxa simultaneously. A 50% majority rule consensus tree was generated from the 106 trees found by this search. Further heuristic searches were conducted based on this initial result, as detailed in the Results section-
Area ciadograms that show the geographic distriiutions of the mtDNA Lineages in
Alberta and California were constructed by plotting the points at which each genotype occurred and enclosing the distribution of a given genotype in a minimum polygon.
These figures were based on the collection sites of previous studies fiom which the samples sequenced in the current study wen obtained.
A model of the evolutionary history of Odocoileus in western North America was constructed fiom estimates of divergence between pairs of nucleotide sequences within and among mtDNA lineages. Estimates of times since divergence were cdculated as the number of pairwise differences observed between any two genotypes, divided by the total number of nucleotide bases analyzed per sequence. The approximate time of divergence of major genotype assemblages was then estimated using published rates for the mtDNA genome as outlined in the Discussion.
The data obtained hmthe deer samples fiom the Magrath, Alberta and Waterton
Lakes National Park, Alberta area were analyzed statistically on a microgeographic scale.
MEGA (Kumar et al., 1993) was used to calculate pairwise haplotype divergences from the 401-nucleotide sequence data for each genotype identified in the region. These divergence vdues and the kquency of occurrence of each genotype in mule deer-and whitetailed deer were then used by the Restriction Enzyme Analysis Package (REAP) program (v.4.0) (McEIroy ef al., 199 1) to calculate the haplotypic or nucleon diversity (h) index for non-selfing populations and the nucleotide diversity (x) index of Nei and
Tajima (198 1) (equations 8.4 and 10.4 of Nei, 1987). The haplotypic diversity (h) index approximates the probability that any two individuais chosen at random fiom within a population will have different haplotypes. A value of0.0 indicates complete fixation of a haplotype, whereas a value of 1.0 indicates that all individuals in the sample have different haplotypes Wei, 1987). The nucleotide diversity (x) index measures the average painvise nucleotide difference between individuals within a sample. This index corrects h for the size of the nucleon studied (Nei, 1987). These iadices were used to estimate genetic heterogeneity within populations, with all mule deer being one population and alI white-tailed deer being another population within the one geographic region. Estimates of nucleotide diversity and nucleotide divergence among populations were also calculated by REAP. The geographic distribution ofmtDNA genotypes within this region was also plotted on a universal transect method (UTM) map. RESULTS
Microgeographic amlysis of deerfiom Magrath and WatertonLakes National Park
Of the 19 white-tailed deer and 42 mule deer fiom the Magrath and Waterton Lakes
National Park area of Alberta for which mtDNA sequence data were obtained, three were
identified as having mtDNA genotypes typically found in the other species. Two mule
deer samples had the EDN genotype, and one white-tailed deer sample had the CYP genotype (Table 1). Based on previous studies (Carr and Hughes, 1993; Hughes and
Carr, 1993), the EDN genotype is typically found only in white-tailed deer, whereas the
CYP genotype has previously only been found in mule deer. One new genotype WC) was found in a single white-tailed deer; all other genotypes were previously reported, though not all genotypes identified by Hughes and Carr (1993) among Odocoileus in western Canada were found in this sampling location (Table 1).
Figure 2 shows the origin of individual deer within southern Alberta at the time of sampling. Map co-ordiaates were available for 55 of the 61 individuals for which sequence data were obtained and these appear on the figure. Most genotypes occur at higher altitudes (the area below the contour Line on the figure). This pattern is more obvious in the mule deer genotypes, with only one representative individual occurring in Table 1 : Mitochondria1 DNA genotypes found in deer from the McIntyre Ranch near Magrath, Alberta, and Waterton Lakes National Park, Albeita Genotype codes are those reported by Hughes and Carr (L993), with the exception of the single new genotype, MAC.
Genotype Number of Deer mule deer white-tailed deer
BNP
PRO CYP
EDN
MVL,
MAC
Total 42 19 Figure 2: Sampling location of white-tailed and mule deer mtDNA genotypes within the southern Alberta colIection site (McIntyre Ranch near Magrath, and Waterton Lakes National Park). Each genotype is represented by a distinct symbol. as indicated in the key. Mule deer genotypes are represented by filled circles and white-tailed deer genotypes are represented by open circles. The region below the contour Line is more mountainous terrain, and the region below the line is predominantly lowland. Values on the abscissa and the ordinate comespond to UTM (universal transect method) A and B coordinates, respectively. A map ofCanada indicates the geographic location of the collection site.
BNP
CYP I
PRO
€DN 0
MVL a
MAC 0 the lowlands (the area above the contour he). The most common mule deer genotype is
BNP (Table 1). Individuals with the white-tailed deer genotype EDN, the most common white-tailed deer genotype in the region, occurred in both the highlands and the Lowlands
(Figure 2).
Mule deer and white-tailed deer fiom the Magrath and Waterton Lakes National Park region of Alberta were gdcaUycharacterized through the cdcuIation of indices of haplotype and nucleotide diversity (Table 3). The pairwise haplotype divergence estimates upon which these cdculations were based appear in Table 2. The haplotype diversity (h) estimates within mule deer and within white-tailed deer are 0.5875 and
0.2902, respectively, which indicate that the probability of two individual deer chosen at random having different genotypes is approximately 59% for mule deer and 29% for white-tailed deer. The nucleotide diversity (z)estimate for within-mule deer was
0.007079, which indicates that any two individual mule deer chosen at random differ by approximately three nucleotide substitutions in the 401-nucleotide region of the cytochrome b gene studied (ie., 401 x 0.007079 = 2.8). The corresponding value for within-white-tailed deer is 0.00252 1, which indicates that any two individual deer of this species chosen at random differ by approximately one nucleotide substitution in the 401- nucleotide region of the cytochrome b gene studied. The within-species estimates of haplotype and nucleotide diversity for mule deer and white-tailed deer indicate that mule Table 2: Painvise haplotype divergence estimates for the mtDNA genotypes identified in mule deer and white-tailed deer from the McIntyre Ranch near Magrath, Alberta, and Waterton Lakes National Park, Alberta. The number of nucleotide substitutions between pairs of genotypes is shown in the upper ma&. The divergence estimates are shown in the lower matrix,
BNP PRO CYP EDN ElVL MAC
BNP - 7 3 5 5 3 PRO 0.01754 - 10 5 6 6 CYP 0.00752 0.02506 - 6 6 4 EDN 0.01253 0.01002 0.01504 - 2 2 MVZ; 0,01253 0.01504 0.01504 0.00501 - 2 MAC 0.00752 0,01504 0.01002 0.00501 0.00501 - Table 3: Indices of haplotype (h) and nucleotide (z)diversity within populations of mule deer and whibtailed deer From the Mchtyre Ranch near Magrath, Alberta, and Waterton Lakes National Park, Alberk
Population Haplotype Diversity Nucleotide Diversity
Mule deer 03875
White-tailed deer 0.2902 deer are geneticdy more variable than white-tailed deer in the Magrath and Waterton
Lakes National Park area of Alberta
New rntDNA sequence genotypes identiTed
Several new mtDNA sequence genotypes were identified in this study among samples collected in westem Caaada Ow new genotype (MAC) was found in a single white- tailed deer from the Magrath and Waterton Lakes National Padc area of Alberta, and a second new genotype (KNP) was obtained fiom a white-tailed deer ern Kootenay
National Park, British Columbia. A white-tailed deer fiom Saskatchewan had a mtDNA genotype (SSK) that was identical to a genotype (CR003)identified in a mule deer collected in southern California. As a result, this genotype (SSK) was coded as a mule deer genotype in phylogenetic analyses. The samples from southern California represent the genetic variation as determined fiom RFLP analysis. The RFLP-genotypes listed in
Table 4 have also been found at several other geographic locations within the United
States, including Montana, Colorado, Utah, Oregon, Arizona and Washington (Cronin,
1991 ; Cronin and BIeich, 1995). the set of samples fiom southern California deer, only one individual had a sequence-genotype reported previously (SLO;Carr and
Hughes, 1993); all other sequence-genotypes are new (Table 4). Two of the RFLP- genotypes @ and 0) have identical sequence-genotypes (CR004). Both samples Table 4 : Correspondence of 401-aucleotide sequence genotype with 307- nucIeotide sequence genotype (Carr and Hughes, 1993; Hughes. 1990; Hughes and Cam. 1993), and with --genotype (Cronin, 1991; Cronin and BIeich, 1995)-
RnP-genotype 307-nucIeotide 40 1-nucleotide cyt b genotype cyt b genotype
CYP
CYP SWN BNP
-
SLO representing RnP-genotype K had the same sequence-genotype (CROOS), and both individuals of RFLP-~~~O~~~~B had a sequence-genotype CR003. For the remaining
RFLP-genotypes (A, L, MyN), the two individuals within each pair could be distinguished &om each other at the DNA sequence level; e.g, the two individuais with
WLP-genotype A had distinct sequence-genotypes, CROO t and CR002.
Comparison ofpreviously reponed mtDNA sequence genorypes cmd new data.
Representative individuals of each mtDNA sequence-genotype reported by Carr and
Hughes (1993) and Hughes and Can (1993) based on 307 nucleotides of sequence data were reanalyzed to give 401 nucleotides of sequence data (ie, an additional 94 nucleotides of sequence at the 5-prime end of the gene fragment were obtained). Some of the mtDNA genotypes identified in deer collected in Canada and in the United States were assigned different three-letter codes when first reported on the basis of 307 nucleotides of sequence; e.g, the genotype TEH reported by Carr and Hughes (1993) genotype ELP reported by Hughes and Carr (1993) have identical sequences. These two genotypes
(TEH and ELP) were also identical when the additional 94 nucleotides of data were considered. In contrast, SAN differs fiom ETX by an RFLP, yet the two genotypes have identical 40 1-nucleotide sequences. BNP and REY are identical to each other on the basis of 401 nucleotides of sequence data, but were distinct on the basis of 307 nucleotides of data. BNP and KIM can be distinguished from each other on the basis of nucleotide substitutions within the additional 94 nucleotides of information, as can SLO and CYP; the 307-nucleotide data set suggested that BNP and KIM, and SLO and CYP were identical in sequence composition. The two individuals representing the KIM genotype (Cam and Hughes, 1993) that were amplified and sequenced had different 401- nucleotide sequences; KIM5 had two nucleotide substitutions not present in KIM4
(Figures 3 and 4). In some cases, substitutions within the additional 94 nucleotides of the
40 1-nucleotide sequence identified variation within genotypes defined by the 307- nucleotide or RFLP genotypes. CROOl and CR002 are variant forms of the RFLP- genotype A / CYP sequence genotype. The 401-nucleotide sequence data shows three substitutions between CROOl and CR002, two substitutions between CR002 and CYP, and one substitution between CROOl and CYP. Similarly, SWN and CR003 are variant forms of the RFLP-genotype B (Figures 3 and 4).
Pattern of nucleotide substitution among Odocoileus mtMsequence genotypes.
The distribution of nucleotide variance in the 40 1-nucleotide data is presented in Table
5. A total of 44 variable sites among Odocoileus genotypes was identified. Most of these sites were two-fold variable, and one was three-fold variable 0,C, A). Most of the substitutions were transitions (89%), with 83% at the third position. This pattern agrees Figure 3 : Mitochondria1 cytochrome b seqwnces of genotypes reported in Odocoileus. DNA sequences represent a 40 1-nucleotide bgment ofthe gene. The inferred amino acid sequence is presented in the first line, and the second line is the reference Odocoileur sequence. Numbers at the ends of the first and second lines indicate the positions in the amino acid and nucleotide sequences, respectively- Dots are used to indicate positions sharing the same nucleotide as the refmnce sequence. Those nucleotide substitutions resulting in amino acid replacements are indicated by asterisks. The mwse (Alcles olces) sequence is presented as an outgroup.
Figure 4 : Character matrix of the variable nucieotide sites used in the phylogenetic analysis of Odoc0iIeu.s refationships. The position of each nucleotide in the original 401-nucleotide SeQuence data (Figure 3) is given in the first three lines, followed by the CMocoiZeus reference seqence. Dots indicate positions where the genotype sequence matches the reference. The moose (Alces alces) is presented as an outgroup sequence. 11 111 111 111 111 111 122 222 222 222 222 222 222 333 333 333 333 333 333 33 12 344 455 666 678 689 900 001 223 344 456 778 900 112 234 455 667 778 899 012 222 445 556 678 888 99 418 045 914 036 981 246 925 895 672 514 502 149 214 692 543 625 140 390 514 681 678 251 473 951 457 03 BNP (md) ATC AAC GAC TTA CCT TAC CCT GCA CCA TTT CCT CCT CTC CTC TCT ACT ACC TTC CTG AAT TCT CTA TCC TCA CGA GC REY (md) ...... KO0 (md) ...... C,, ...... KIM5 hd) GC, .,. AG. .,.,.o..,..s.,*,,..,*...... ,.,....,.....*...... A, .. KIM4 (md) GC, .., A,, ...,.....,.....,,.,.,...... ,...... ,...... CR006 (md) ,..,,,... C., *....,.*.....*...,.....,..T *...... T,, ...... ELP (md) ,.,,.. ,a. c., ...... ,.,....,....,...... T., ...... TBH (md) ...... a. C...... ,*...,..,,,...... ,....T,, ...... ,... CROO4 (md) ,~~~~~..~.,~~...... ~~~~~~~~.~~~~~..,,~....,T...... WTX (md) ... .C...... T...... CR007 hd) .*,,..,.* .C. T,C ,*,*.*..*.,,,*...,...,,*... .T* ...I...... ,.....,...... , CROl1 (mdl ,..,.....,.. T...... ,...... , .TC ,...,.,...... ,...... SM) (md) ...... , T,. ..,,....,.., .,C .,...... TC ....~,~.~~~~~..~~~...~~.....~~~~ CR002 (md) ,.*..,...... T,, O**O.,... T,, .,c .,,,,*,,...* .Tc ...... CROOS (md) **...... * T,. *I...,.*,,,, .,C ,,,I,.,,,.*, .TC ...... ,,G ..... CY P (md) ...... **C ,..,..,...,. ,TC ...... ,...... I CROO 1 (md) ,,....,,..,,*..*..,.I.**...,,c .*C ,.....,.. *TC ,**...... ,,,...... I....,.. CROO 8 Imd) ,-9 a,, ,I, ,C, ,,, ,,, 9,- *.e I,, 9.. a,, ,,C ,,. .,, ,,, ,,, ,,, 90, ,,, ,,, ,,, ,,, ,, CROO 9 (mdl ..,,.. ,Q, ...... ,.,,,,..,,*...... ,*..,,*c .,.,...,,,.,..*...... I....,... SWN (md) ,.....,.. ,C, ...... ~..,...~.....,...... ,,...... ,.TC...... T PRO (md) .,. .,, ,Q, ,C, ,,, A,, ,,, ,,, ,,, ,,. ,.. ,,, ,,, ,,. ,,, ,,, ,,, C,, TC. .,. ,,, ,., ,,, .,, ,,, ,T COY hd) ,,,... ,a, ,C. ..*...,*..*...v.*.....v,,..*.,.,*...... TC. ..I..,,....,.., ,A, ,T CR003 (md) .*l..I ,a. .c, ..,...,,..,,,.,,,*.*...,,,..,.,.,,...O,...TC. ....I...... ,T SSK (md) ...... ,a, ,C...... TC, ...... ,T SAN (wt) ...*.....*C* ,..**.....*...... ,,,..C ...... C, ...... T ETX (~t).....I... .C...... C ...... C...... I...... ,... ,T LOB (wt) ...... I .CI ...... I.... I.C .I.... ,,C l..~...... Ct ...... ,T GRA (wt) .I.1..... ,C, ...... I..T.. ,.C ,,...... C ....,,... ,C, ...... I....1...... ,T MVL (wt) lll...... C. ....I...... , ..G ...... ,C ...... ,C, ...... ,T EDN (~t)...... *G1 .CI ....e.....e...... C ...I...... CI ...... ,T MP (WU ..,.,. +a. ,c. ,...... ,,,,c ~~~~...~~...,.~,~,.~~..,~~...~,T MAC (wt) ...... *.....*...... C ...... C* .ll....,....,...... T HOP (bt) .I..v.q.. *.T **. *GI **.a** T**,CC *TC .I) TCT Ice *...l.... CC, TI, GtC .*CTI, ,.. (T, T,, I, AKB (bt) ... C., ... ,.T .,C .Q...... T., .CA .TC ... TCT .Co ...... CC, T., G.C .,C T,, ... ,T, T,, .. AWES ..T ,,T ,,T ,,, ,T, ,,T QTC ATG .T. CAA .., TTC ,CT TAT C,. GAC CAT CCT T,A .C, ,TC ,CG CTT A,, .,T AT Table 5: Frequencies and distributions of variant nuckotides in Odocoileus - within the 40 1-nucleotide cytochrome b sequence data. A total of-@ variable sites w& identified, one of which is three-fold variable.
Codon Position
Fit Second Third Totai
Transitions 4 3 33 40
Transversions 1 I 3 5 with that reported in a previous study of Odocoileur in western Canada (Hughes, 1990).
Transversions accounted for 1I% of the variance, Seven of the nucleotide substitutions result in amino acid substitutions (Figure 3).
Phylogenetic analyses of rntDNA sequence genotypes.
The neighbor-joining tree constructed fiom the Kimura 2-parameter distance matrix
(Figure 5) shows the presence of several groups of seqyence genotypes. The black-tailed deer genotypes (HOP and AKB) clustered together and are the most divergent of the
0docoiZeu.s genotypes. Most of the mule deer genotypes are distributed across the middle portion of the tree, with the exception of a group of five genotypes (SWNy CGY, PRO,
CR003, and SSIC) that is located within the white-tailed deer genotype cluster (LOB,
GRA, ETX, MVL, SAN,MAC, EDN,and KNP). The mule deer genotypes CR008 and
CR009 are also part of the white-tailed deer cluster. The neighbor-joining tree constructed using synonymous sites only (Figure 6) is almost identical to the tree that results fiom analysis of all sites simultaneously; the same major clusters of genotypes are present. In contrast, the neighbor-joining tree constructed using noasynonymous sites only (Figure 7) shows essentially no resolution of the phylogenetic relationships among
0docoiZeu.s genotypes.
While the results of the maximum-likelihood quartet puzzling analysis (Figure 8) shows Figure 5: Neighbor-joining tree constructed fkom Kimura-2 parameter pairwise distance estimates (transitions + transversions). The moose (Alces alces) w& used as an outgroup.
Figure 6: Neighbor-joining tree constructed using synonymous substitutions with a Jukes and Cantor correction. The moose . (Alces alces) was used as an outgroup.
Figure 7; Neighbor-joining tree constructed using nonsynonymous substhitions with a Jukes and Cantor correction- The moose (Alces dces) was used as an outgroup. SSK CR008 SLO
SWN CYP .
COY Figure 8: Quartet puzzling tree of the phylogenetic relationships among Udocoilacs mtDNA sequence genotypes. The moose (Alces alces) was used as an outgroupP Reliability percentages are given for the internal nodes. less resolution than the neighbor-joining analysis of all sites, some of the same clusters of genotypes are present in both trees. The set of five mule deer genotypes (SWN, PRO,
CGY, CR003, SSK) is present as a group in the quartet puzzling tree as well as the neighbor-joining treeyfor example. This group was reconstructed during 67% of the
1000 puzzling steps.
The complete 401-nucleotide sequence data set included variants at 44 positions among the Odocoi[eus genotypes; cladistic analysis was performed on this subset of the data
(Figure 4). Initial analysis of the original character matrix identified 106 identical minimum-length trees of length LOO and consistency index 0.8000. The 50% majority rule consensus tree of the 106 minimum-length trees (Figure 9) shows a comparable degree of resolution to the neighbor-joining tree (Figure 5). The two black-tailed deer genotypes (HOP and AKB) form a distinct clade relative to the other Odocoileus sequence genotypes. The same set of mule deer genotypes (SWN, PRO, CGY, CR003, and SSK) appear within the white-tailed deer genotype clade (SAN, ETX, LOB,GRA,
MVL, MAC, EDN, and KNP), as do the mule deer genotypes CROOB and CR009. The relationships among all genotypes were not fully resolved by the initial heuristic search.
As a result, two exploratory searches were conducted.
The distribution of the variation among Odocoileus genotypes (Figure 4) suggests the presence of four main groups. Black-tailed deer (HOP and AKB) are clearly a separate Figure 9: Maximum parsimony 50% majority nrle consensus tree of the 106 minimum length trees recovered by a heuristic search of the unmodified mtDNA sequence genotype data. The percentage of the trees in which groupings occurred appear on the tree. The moose (Alces akes) was used as an outgroup- Majority rule IBNP(md) REV(md) KOO(md)
WTX(md) C R007(md) CROl l(md) 11 00 -SLO(md) CROO2(md) 10%- CR005(md) I001 CY P(md) CR001(md) C R008(md)
,-, ,-, SWN(md)
I
100
- 100 HOP(bt) 1 AKB(bt) ALCES group since the two genotypes shma number of substitutions unique to tbat group.
White-tailed deer genotypes (s& ETX, LOB, GRA, MVL, EDN, KNP, and MAC) also
share many of the same substitutions within their group; most genotypes share the same
substitutions at sites 63,243,291, and 393. In terms of the mule deugenotypes, there
appear to be at least two groups, one of which shares three substitutions with the white-
tailed deer genotypes (SWN, PRO, CGY, CR003, SSK; sites 63,291,393), and the other
(all remaining mule deer genotypes) of which shares only one substitution with the white- tailed deer genotypes (site 243). The presence of a T <-> C transition at site 243 in all white-tailed deer genotypes and some mule deer genotypes appears to be the result of convergent evolution (homopiasy), since it is the only piece of evidence grouping these mule deer genotypes with the white-tailed deer genotypes. The noisiness of this site obscures the phylogenetic signal of the data, as it causes these mule deer genotypes to sometimes group with the white-tailed deer genotypes even though there are more pieces of evidence grouping these mule deer genotypes with the remaining mule deer genotypes.
In order to overcome the problems associated with homoplasy, this character state CC') at site 243) was recoded as 'A' in all mule deer genotypes that had a %. An initial heuristic search carried out on white-tailed deer genotypes produced a single minimum- length tree of length 6, while the equivalent analysis of mule deer genotypes identified a single minimum-length tree of length 3 1. The mule deer tree was used as a backbone constraint tree in a second round of analysis that included all white-tailed deer genotypes. It is possible to search under topologicd constraints to restrict the set of trees retained by a heuristic search. A backbone constraint tree contains a subset of the study taxa When searching under topological constraints, only those trees that share the topology of the constraint tree will be retained when the constraint tree is used as a filter-
A trial tree is compatible with the constrai~ttree only if pruning those taxa on the trial tree that are not present on the backbone tree results in a topology identical to the backbone (SwoEord and Begle, 1993). This type of constraint keforces a relative topology, since other taxa may be added at any point on the backbone tree, providing the backbone is not violated (Swofford and Begle, 1993). The second round of analysis identified 77 minimum-length trees of length 39. When the mule deer constraint tree was used as a filter, five trees were retained that were compatible with the constraint tree, and were saved for later use as constraint trees. The two black-tailed deer genotypes were then added to the data set for analysis. The heuristic search identified 162 minimum- length trees of length 61. When the mule deer genotype constraint filter was applied, 14 trees were retained; independent application of the [mule deer + white-tailed deer] constraint filters retained between 3 and 11 trees per constraint tree. The fact that a single tree was not recovered suggested that there may be other sites within the data set that are phylogenetidy noisy. Further inspection ofthe data set suggested that the distribution of character state 'G' at
character eight (nucleotide 5 1 of the original data set) was also the result of convergent
mutations in mule deer and white-tailed deer genotypes- Character eight was therefore
excluded fiom a second exploratory analysis. The complete (or partial, as in the case of
recnding a character) exclusion of clearly unreIiable data from an analysis is an
acceptable solution to the problem of homoplasy (Swofford et aL, 1996). The same
search criteria were used in phylogenetic axdysis of the modified character data matrix.
The exclusion of character eight caused some genotypes to collapse onto each other; for
example, SWN and CR003/SSK become identical (Figure 4). The second modification
(i.e., exclusion of character eight) was used in all subsequent analysis.
Analysis of only the mule deer genotypes identified a single minimum-length tree of
length 27 (Figure lo), and independent analysis of only the white-tailed deer genotypes
identified a single minimum-length tree of length 5 (Figure 11). When the mule deer
genotypes and the white-tailed deer genotypes were dysedtogether in the same heuristic search, a single minimum-length tree of length 32 (Figure 12) was identsed-
The black-tailed deer genotypes were then added to the analysis; this identified only two minimum-length trees, each requiring 54 steps (Figures 13 and 14). These trees differ with respect to the position of the black-tailed deer genotypes relative to the mule deer and the white-tailed deer genotypes. Constraint trees were not used as filters in this Figure 10 : Unrooted minimum-length network of phy logenetic relationships among mule deer mitochondria1 DNA genotypes in North America. Filled circles tepresent genotypes; bars indicate the number of nucleotide substitutions inferred to occur along each branch of the network,
Figure 1 1 : Unrooted minimum-Length network of phylogenetic relationships among white-tailed deer mitochondria1 DNA genotypes in North America Open circles represent genotypes; bars indicate the number of nucleotide substitutions infdto occur along each branch of the network
Figure 12 : Unrooted minimum-length network of phylogenetic relationships among mule deer and white-tailed deer mitochondrial DNA genotypes in North America Mule deer genotypes are represented by filled circles and white-tailed deer genotypes are represented by open circles. Bars indicate the number of nucieotide substitutions inferred to occur along each branch of the network.
Figure 13: One of two unrooted minimum-length networks of phylogenetic relationships among mule deer, white-tailed deer, and black-tailed deer mitochondrid DNA genotypes in North America. Mule deer genotypes are represented by filled circles, white-tailed deer genotypes are represented by open circles, and black-tailed deer genotypes are represented by circles with diagonal lines. Bars indicate the number of nucleotide substitutions inferred to occur along each branch of the netwok The number ofnudeotide substitutions differentiating the black-tailed deer genotypes hmthe mule deer and white-tailed deer genotypes is indicated by numerals-
Figure 14 : The second of two unrooted minimum-length networks of phylogenetic reIationships among mde deer, white-tailed deer, and black-tailed deer mitochondrial DNA genotypes in North America Mule deer genotypes are represented by filled circles, white-tailed deer genotypes are represented by open circles, and black-tailed deer genotypes are represented by circles with diagonal lines. Bars indicate the number of nucleotide substitutions infidto occur along each branch of the network The number of nucleotide substitutions differentiating the black-tailed deer genotypes from the mule deer and white-tailed deer genotypes is indicated by numerals. This network differs fiom that presented as Figure 13 in terms of the position of the black-tailed deer genotypes relative to the mule deer and the white- tailed deer genotypes-
andysis.
Phylogenetic analyses identified four major assemblages of ~docofieusmtDNA
genotype: (1) black-tailed deer genotypes (AKB and HOP), (2) white-tailed deer
genotypes (XNP, EDN, MVL, MAC, LOB, and GRA), (3) a group of mule deer
genotypes predominantly found in western Canada (PRO,SWN, and CGY), and (4) a
group of mule deer genotypes found predominantly in the westem United States (the
remaining 17 mule deer genotypes) (Figures 13 and 14).
The geographic organization of the genotype assemblages can be summarized as follows: assemblage (1) western Noah American black-tailed deer, (2) western North
American mule deer, (3) western Canadian mule deer, and (4) western North American white-tailed deer. The minimum and maximum pairwise nucleotide sequence divergence estimates within and among genotype assemblages are presented in Table 6. The maximum interspecific sequence divergence is 698% between western North American white-tailed deer and black-tailed deer, while the maximum intraspecific sequence divergence is 6.73% between western North American mule deer and black-tailed deer.
The approximate geographic distributions of mtDNA genotypes in California and
Alberta (Figures 15, 16 and 17) indicate that there is a considerable degree of geographic structure to the sequence genotypes. Within both 0.hemionur and 0.virginianus, some genotypes are widespread while others are more localized. For example, among white- Table 6: Minimum and maximum pairwise nucleotide sequence divergences within and among Odocoileus mtDNA genotype assemblages. The estimated sequence divergences were calculated as (number of painvise differences) / 40 1 .
Genotype assemblage
1 Genotype assemblage
I. western North American black-tailed deer 0.75%
2, western North American mule deer 4.75% - 6.73% 0,25% - 2SO% 8 3. western Canadian mule deer 5.49% - 6.48% 1 ,OO% - 2.99% 0.25% - 0.75%
4. western North American white-tailed deer 5.74% - 6.98% 0.50% - 2,74% OSO% - 1.75% 0,25% - 1 ,OO% Figure 15: Approximate distribution of mule deer (Odocoifeushemionur) mtDNA genotypes in California The extent of occurrence of a given genotype is bounded by a minimum polygon. Numbers are used to represent genotypes as follows: 1 HOP 2 TEH 3 BNP 4 SLO 5 CROO1 6 CR003 7 CROOS 8 CR006 9 CROf 1 10 CR007 11 CR002 12 CR009 13 CROW 14 CROOS
Figure 16: Approximate distribution of mule deer (Odocoileus hemionm) mtDNA genotypes in Alberta, The extent of occurrence of a given genotype is bounded by a minirnum polygon. Numbers are used to represent genotypes as follows: 1 sw 2 PRO 3 BNP 4 TEH 5 KO0 6 CGY 7 CYP
Figure 17: Approximate di-iution of white-tailed deer (Odocoileur virgr'niamcs) mtDNA genotypes in Alberta. The extent of occumnce of a given genotype is bounded by a minimum polygon Numbers are used to represent genotypes as follows: I EDN 2 m 3 KNP 4 MAC tailed deer genotypes in Alberta, EDN is the most widespread, while MAC and KNP have
been found at one location each (Figure 12). DISCUSSION
PaZaeontol~~calrecord of Odocoileus.
Fossil material attributed to Odocoileus first appeared 3.5 x 106 years ago. Material
classified as 0.virginiamcs dates to 3.2 x lo6 years ago, and 0.hemiomrs material first
appears in the Irvingtonian (0.7 - 19 x lo6) (Kurtt2n and Anderson, 1980). Fossils
attributed to 0.virginimrus are found in Blancan Florida and Texas deposits. The North
American Land Mammal Ages are defined by fadassemblages, and have been
independently dated using radiometric and palaeomagnetic techniques. There is some
variation in the start and end dates of palaeontological ages, depending on which
authority is consulted; for example, Martin and Barnosky (1993) state that the Blancan
extended &om 4.8 - 4.0 x LO6 years ago to 1.9 x lo6 years ago, while Kurt& and
Anderson (1980) state that the Blmcan began 3.5 x 106 years ago. By the Irvingtonian,
fossils attributed to white-tailed deer were present in several of the central and eastern
states, and by the Rancholabrean were very common and widespread. Most of these records are from central and eastern North America, though there are also several western fossils of OdocoiIeus that may be 0. virginiunus. Fossils recognimble as 0.hemionus
are limited to about 15 sites in western North America; several other western fossils of
Odocoileus, most ofwhich date to the Rancholabrean, may also be 0.hemioms (Kurt&.~ and Anderson, 1980). The palaeontological record as conventionally interpreted would thus suggest that white-tailed deer are ancestral to both mule deer and black-tailed deer. -
Evolutionmy history of Odocoileus based on morphoZogrgrcuZand mfDNA sequence data
Phylogenetic analysis of mtDNA sequence genotypes suggests a contrasting evolutionary history of Odocoi~euseusThe sequence data suggest that black-tailed deer are ancestral to both mule deer and white-tailed deer (Figures 5, 13 and 14); interpretation of morphologid variation within 0.hemioms also suggests that black-tailed deer are ancestral to mule deer, and possibly all Odocoileus species (Cam and Hughes, 1993;
Geist, 198 1). Among the criteria that segregate ancestral fiom derived forms of ungulates, Geist states that more highly evolved forms are typically larger and have a more distinctly patterned pelage, a larger rump patch, and a shorter tail. The adaptive trend involves a shift fiom forest to open habitat, fiom low latitudes and altitudes to higher ones, and &om old glacial refugia into more recently glaciated zones. On the basis of these criteria, the ancestral subspecies of 0. hemionus is the black-tailed deer that inhabits the Alaskan coast, the Sitka deer (0.?i sitrtenis), and the most derived subspecies is the Rocky Mountain mule deer (0.h hemionus) (Geisf 1981).
The maximum sequence divergence between a black-tailed deer genotype (Am) and any other Odocoileus genotype (the white-tailed deer genotype GRA) is 698% (Table 7). On the basis of RFLP data, Brown et al. (1979) estimated the rate of sequence divergence
for the entire mtDNA molecule as 2% per million years per pair of lineages, or 20 x
substitutions per nucleotide site per year per pair of Lineages. Nei (1985, cited in Hart1
and Clark, 1989) estimated the rate as 0.71% per million years, or 7.1 x substitutions
per nucleotide site per year per pair of lineages, a three-fold difference- Previous work
on Odocoileus (Carr and Hughes* 1993) indicates that the portion of the cytochrome 6
gene under study evolves slightly faster than the entire molecule: they calculated the
sequence divergence between black-tailed deer and white-tailed deer as 7.5% for a 307bp
hgment of the cytochrome b gene, whereas the divergence between the same two
Odocoileus genotypes calculated fiom RFLP data for the whole mtDNA molecule was
6.0%. Using the corrected cytochrome b divergence rate of Carr et al. (1995). which is based on Nei's (1985) divergence rate and corrects for the faster rate of evolution of the cytochrome b gene, the 6.98% sequence divergence between the two most divergent hemionus and virginimtus genotypes (AKB and GRA, respectively) suggests a divergence time of ca. 7.9 x lo6 years ago. The calculation fhm Brown et aL1s(1979) rate, corrected for cytochrome 6, used by Cam and Hughes (1993) suggests a divergence time of ca 2.8 x lo6 years ago. The first estimate is incompatible with the palaeontological record, since
it antedates the appearance of the genus, or even New World deer, in the fossil record.
The more recent estimate obtained on the basis of Brown et al.'s (1979) rate is compatible with the fossil record; all subsequent estimates of time since divergence are based on
Brown et uL's (1979) rate.
The earliest estimated time since divergence between a mule deer genotype and a white-tailed deer genotype is considerably more recent; on the basis of Brown et aL1s figure, the 2.74% sequence divergence between KIM5 and GRA (Table 6) cornsponds to a divergence time of CP 1.1 x 106 years ago. Mule deer and black-tailed deer genotypes diverged from each other a maximum of ca 2.7 x lo6years ago (maximum sequence divergence of6.73% between AKB and CR005; Table 6). Based on these estimates, the black-tailed deer genotypes represent the oldest 0docoiIeu.s mtDNA lineage in North America The mule deer mtDNA lineage diverged from the black-tailed deer, and the white-tailed deer mtDNA lineage subsequently diverged fiom the mule deer lineage. This general order is indicated by the brauchkg pattern shown in the phylogenetic networks (Figures 13 and 14).
The lack of concordance between the evolutionary patterns among Odocoileus lineages as determined fiom mtDNA sequence analysis and fiom the fossil record may be due in part to the fact that many of the Odocoileus fossils have not been classified to a particular species, let alone subspecies. Species designation assigned to a given fossil may also be iduenced by the collection site: Odocoileus specimens fiom eastern or centrai North
America may be assumed to be 0. virginimnrs simply for geographic reasons (C. Dailey, pers. comm. to S. Carr). KurtCn and Anderson (1980) suggest that Pleistocene
0docoiieu.s may require revision.
The evolutionary history of Odocoiletls as indicated by molecular data is complex.
Previous studies have demonstrated a discordance of nuclear and mitochondria1 genetic differentiation: mtDNA sequence divergence was found to be high between conspecific mule and black-tailed deer and low between the two speciesf 0.hemiorms and 0- virginianm, whereas the relationships indicated by alfozyme data (McClymont et al.,
1982; Stubblefield et al., 1986) were in agreement with the classical species and subspecies designations (Carr and Hughes, 1993; Cronin, 1991; Cronin et al., 1988). The current study indicates the same pattern of high intraspecific sequence divergence and low interspecific sequence divergence among mtDNA genotypes: the maximum intraspecific divergence for 0. hemionus genotypes (6.73% between AKB and CR005) exceeds the maximum intraspecific divergence for 0. virgznimus genotypes (1.00% between GRA and KNP) in western North America Uable 6). RFLP analysis of mtDNA among white-tailed deer in the southeastern United States indicates that the level of nucleotide sequence divergence among southeastem white-tailed deer RFLP-genotypes
(maximum of 2.8 1%) is greater than that found in white-tailed deer populations fiom other regions of North America (Ellsworth et al., 1994). Direct DNA sequence analysis of deer fiom the southeastern United States would be expected to show a similar pattern as the RFLP analysis of DNA-
The phylogenetic networks show four main assemblages of rntDNA genotypes: (1) black-tailed deer genotypes (AKB and HOP), (2) white-tailed deer genotypes m,
EDN, MVL, MAC, LOB, and GRA), (3) a group of mule deer genotypes predominantly found in western Canada (PRO,SWN, and CGY), and (4) a group of mule deer genotypes found predominandy in the western United States (the remaining 17 mule deer genotypes) (Figure 13). The placement of the white-tailed deer assemblage within the network is of particular interest: the white-tailed genotypes phylogenetically separate the two mule deer assemblages. Inspection of the sequence data indicates that the white- tailed deer genotypes are quite similar to the mule deer genotypes eom western Canada
(Figure 4). In order to estimate approximate times since divergence among the major assemblages, the most widespread genotype within each assemblage was used as a representative sequence, and times of divergences were estimated using Brown et aL's
(1979) 2% per million years per pair of lineages rate of sequence divergence, corrected for the cytochrome b gene.
A model for OdocoiZew mtDNA evolution suggested by the current study is presented as Figure 18. According to the relationships shown in the model, the black-tailed deer diverged fiom the ancestral stock somewhat more than 2 MYBP. The next divergence event did not occur until about I MYBP; this event separates the southeastern white- Figure 18: Model of evolutionary patterns among Odocoileus mtDNA assemblages in North America Mule deer genotype assemblages are represented by filled circles, white-tailed deer genotype assemblages by open circles, and the black-Wed deer assemblage by a circle with diagonal Lines. Branch points indicate approximate sequence divergence date estimates as calculated hmmtDNA sequence data No sequence data were available for the southeastern USAand southern Florida assemblages; estimated divergence times for these assemblages are based on RFLP analysis of the mitochondrid genome as reported by EUsworth et (11. (1994).
tailed deer fiom all Odocoileus genotypes currently represented in western North
America On the basis of a recent RFLP &alysis of mtDNA variation among white-
Weddeer in the southeastern United States, there is a southern Florida assemblage of
white-tailed deer RFLP-genotypes that can be distinguished cladistically fiom other
southeastern RnP-genotypes (Ellsworth et al., 1994). Divergence of this southern
Florida assemblage is estimated to have occurred somewhat less than 1 MYBP. The divergence at about the same time or slightly before separates the western United States mule deer assemblage fbm the mule deer assemblage found in western Canada and the white-tailed deer found in westem North America. The most recent divergence suggested by mtDNA sequence data occurred within the last 0.50 x lo6 MYBP and separates
Canadian mule deer genotypes fiom western North American white-tailed deer genotypes.
Glacial reficgia, stochastic lineage sorting and introgressiw hybridization of mtDNA.
Two main hypotheses that have been suggested as explanations for the discordance between nuclear and mitochondrid genetic differentiation in Odocoiletrs are stochastic lineage sorting ofmtDNA, and extensive introgressive hybridization of mtDNA between the two species (Cam and Hughes, 1993; Cronin, 1991). These processes can also explain . the evolutionary patterns observed over the geographic range studied here. Given the extreme divergence of the black-tailed deer assemblage relative to the other OdocoiIeus assemblag&, it can be hypothesized that the black-tailed deer genotypes represent an ancient mtDNA Lineage that has persisted over evolutionary time. The estimated time of separation of this Lineage from other lineages antedates the separation of the species (Cam and Hughes, 1993). Through the process of lineage sorting, aLl other black-tailed deer rntDNA lineages would have been ehnhted, Leaving only the current lineage represented by AKB and HOP. The persistence of this lineage may be attributable to its isolation fkom other 0docoiIeu.s Lineages during the Glaciation. Black-tailed deer probably occupied a coastal fefirgium in the Pacific Northwest, while mule deer may have occupied a rehgium in the southwest (Cronin, 199 1). Contemporary populations of black-tailed deer are largely coastal; the distribution of Columbian black-tailed deer (0. h. columbianus) extends fiom central California to southern British Columbia, and is contiguous with that of Sitka black-tailed deer (0.h. sitkenrs), which extends northward along the coast of British Columbia to southeastern Alaska (Carr and Hughes, 1993).
The high degree of genetic variation in terms of both numbers of distinct mtDNA genotypes and sequence divergences among genotypes within all observed mule deer genotypes suggests that the mule deer also represent an old Lineage that survived the Ice
Age in one or more refugia The three large regions that sewed as the primary terrestrial refbgia during the Wisconsin Glaciation were mid-latitude North America (i-e., the land south of the Laurentide and Cordilleran ice Sheets), Beringia (present-day Yukon aad
Alaska, plus the beds of the Bering and Chuckchi Seas), and ihe coastal plains east of the ice sheet (the currently submerged continental shelf off New England and Atlantic
Canada) (Fielou, 1991; Figure 19). Mule deer may have occupied a southwestern refugium (Cronin, 1991) or Beringia The promeration of mule deer genotypes would have occumd while the species occupied its refbgium and as it extended its range to the east and south as the ice sheets receeded
White-tailed deer would also have occupied refugia during the Ice Age. While a
Florida refigium has been suggested for southeastern populations of 0.virginimtus
(Ellsworth et a[., 1994), the central and western populations may have occupied a southwestern refbgium or may have moved northwards to Be~giaas the ice sheets advanced. If white-tailed deer and mule deer occupied the same rekgia, or at least the same general area, at the same time in palaeontological history, a comparable level of genetic polymorphism would be expected in both species. The results of mtDNA sequence anaIysis suggest the contrary: 20 distinct mule deer mtDNA genotypes have been identified in westem North America, while only 7 white-tailed deer mtDNA genotypes have been identified in the same broad geographical region. The lack of extensive polymorphism in westem white-tailed deer may indicate that this mtDNA genotype assemblage is of more recent origin thau either the western mule deer or the Figure 19: Extent of ice coverage over Northern North America during the Wisconsin Glaciation. The maps show the increase in the relative amount of ice-bland available for colonization by organisms at 18,000 years ago (A), 13,000 years ago (B), 10,000 years ago (C),and 7,000 years ago @). These maps are modified &om Figures 1.2 - 1.5 in Pielou (1991). southeastern white-tailed deer. Given that both species were well established before the onset of the Wisconsin Glaciation, and that then are severd western fossils of
Odocoileus that may be 0.virginanus (Kurt611and Anderson, 1980), it is reasonable to assume that there was a western white-tailed deer mtDNA lineage. The fate of this lineage may be accounted for by an episode of ancient hybridization between mule deer and white-tailed deer, with iutrogrpssion of mtDNA from mufe deer into white-tailed deer. Such a scenario would effectively replace the white-tailed deer mtDNA lineage in western North America with mule deer mtDNA. The divergence date estimate suggests that such an event occumd fairly recently in palaeontological time (Figure 18).
Gene trees and species trees.
While the hypotheses of stochastic lineage sorting ofmtDNA and of extensive introgressive hybridization of mtDNA between the two species of Qdocoileus can be used to explain the observed discordance between nuclear and mitochondrid genetic differentiation in these deer, it is important to realize that the phylogenetic tree generated from mtDNA sequence data represents a gene tree, and may differ fiom the species tree.
A species tree represents the actual evolutionary pathway of a group of species, whereas a gene tree is a phylogenetic tree constructed fiom DNA sequences for a single gene in each of the species involved &i, 1997; Pamilo and Nei, 1988). Gene trees and species trees can differ as a result of such factors as retention of ancestral polymorphisms and hybridization events. If thetime of divergence between the genes is roughly equal to the time of divergence between populations, the gene tree and the species tree will show the same topology.
In the case of Odocoiieus mtDNA, the obsemation that mule deer are more divergent fiom the conspecific black-tailed deer than fiom the congeneric white-tailed deer may reflect the difference between a gene tree and a species tree. Classical species and subspecies designations are based on morphological and ecological characteristics of the biological entities called black-tailed deer, mule deer, and white-tailed deer. Analysis of alloyme data suggests an evolutionary history consistent with the classical taxonomy
(Cronin, 199 1; McClymont et a[., 1982; Stubblefield et a[-,1986). In contrast to alloyme analysis in which several loci are examined in the same study, mtDNA studies are effectively based on a single locus, a consequence of the nonrecombining mode of inheritence of the mtDNA genome (Moritz et al.,1987). As a result of this mode of inheritance of mtDNA, the effects of hybridization events and lineage sorting of ancestral polymorphism are potentially retained through subsequent generations, and may result in the gene tree recovering a different evolutionary history than the species tree. The phylogenetic trees presented in the current study represent the evolutionary history of the mtDNA sequence genotypes only; the fact remains that black-tailed deer, mule deer, and white-tailed deer are distinct biological entities, regardless of the observed discordance between nuclear and mitochondrial DNA evidence.
Phylogeography of Odocoileus.
The data obtained fkom Odocoileus samples from southern Alberta can be used to illustrate some of the principles of phylogeography on a miaogeographic scale.
Figure 2 indicates the relative distributions ofmtDNA sequence genotypes identified among individuals hmthe McIntyre Ranch near Magrath, and Waterton Lakes National
Park. Since most individuals were sampled as fawns, the geographic co-ordioates corresponded to localities within the respective maternal ranges. In some cases, fawns suspected of being siblings were sampled. As expected from the mode of maternal inheritance, all putative siblings shared the same mtDNA sequence genotype. There are two basic social units among Odocoileus species: doe (or family) groups, and buck groups (Marchinton anci Hirth, 1984). A doe group consists of an adult doe plus her fawn and offspring fiom the previobs year or years. While males typically disperse far fiom the mated range as yearlings (12 - 24 months) to establish solitary residence on new ranges, most female offspring remain in or near the ranges of their mothers (Bunnell and
Harestad, 1983; Hawkins and Klimstra, 1970; Nelson and Mech, 1984; Tierson el aI. ,
1985). Studies have indicated that female dispersal is very limited in comparison to that of males (Nelson and Mech, 1992).
Since dispersal is male-biased, with fedesexhibiting philopatric behaviom, distributional patterns of rntDNA variation would be expected to show geographic structure. Figure 2 indicates clustering of genotypes within the microgeographic study site. Similarly, qualitative analysis of the geographic distniutions of mtDNA genotypes on a broader scale reveals a Ievel of geographic structure, with some genotypes being found at only one location and other genotypes being quite widespread (Figures 15, 16 and 17).
Site fidelity will have its most profound effkct on phylogeograpbic structure if it involves fidelity to reproductive site, since it is there that any exchange of genetic material will occur. For example, a phylogeographic analysis of mtDNA control-region sequence variation in dunlins (Calihs alpina) indicated that this species of migratory shorebird exhibits a strongly subdivided genetic population structure that is being maintained by philopatric behaviour (Wenink et al., 1996). In contrast, mtDNA variation in song sparrows (Melospizu melodia) was not geographically structured, despite geographic variation in size and plumage colour across its North American range (Zink and Dittmann, 1993). RFLP analysis of mtDNA in humpback whales (Megapteru novaeangliae) revealed a strong segregation of mtDNA haplotypes among subpopulations as well as populations fiom the North Pacific and western North Atlantic Ocean populations. This pattern of geographic structure was interpreted to be the result of maternally directed fideIity to migratory sites (Baker et al., 1990). A study of another marine mammal, the grey seal (Kalichoerus -us), revealed a pronounced segregation between western North Atlantic and eastern North Atlantic groups, as indicated by the absence of shared haplotypes (Boskovic et aL, 1996).
Some of the recent studies of Odkieusexamine the genetic aspects of its evolution.
Ellsworth et al. (1994) specifically studied the historical biogeography and contemporary patterns of mtDNA variation in white-tailed deer. The study site was restricted to the southeastern United States, and revealed three main pups of RFLP- genotypes that were geographically oriented and spatially concordant with patterns observed in unrelated species. Most of the other Odocuileur studies focus on the genetic interactions between the two species &or their subspecies in tenns of hybridintion and genetic introgression. The current work examines the phylogenetic relationships among
Odocoileus genotypes over a broader geographic range. While independent studies have been carried out on deer populations throughout their species ranges, these studies have tended to be geographically localized Most used RFLP or dlozyme analysis of genetic variation as a means of studying population dynamics. Given the difference in resolution between RFLP analysis and mtDNA sequence analysis, as indicated by the current aualysis of a 40 1 base pair fhgment of the cytochrome b gene, the data £?om other analyses cannot be directly incorporated. This egectively reduces the extent of geographic sampling of 0docoiIm.s populations to those regions for which tissue and/or
DNA samples were available for DNA sequence analysis Conclusions made on the basis of sequence data may not reflect the phyiogeographic patterns that wodd be observed if the entire Odocoileus range had been sampled. Samples hmsoutheastern U.S.A- for
DNA sequence analysis wodd be particularly valuable, given the phylogenetic patterns among white-tailed deer reported here and by Ellsworth et al. (1994); if all western North
America white-tailed deer mtDNA genotypes have effively been replaced by mule deer mtDNA genotypes via introgressive hybridization, then attention should be tumed eastward in order to gain insight into the white-tailed deer evolutionary history as indicated by mtDNA REFERENCES
Attardi, G. 1985. Animal mitochondrial DNA: an extreme example of genetic economy. Int. Rev. Cytol. 9393-145.
Avise, J. C. 1991. Matriarchal liberation. Nature 352: 192.
Avise, I. C. 1994. Molecular Markers, Natural History and Evolution. Chapman and Hall, Irrc-, New York.
Avise, J. C., I. Arnold, R. M. Ball, E. Bermingham, T. L?mb, J. E. Neigel, C. A. Reeb, and N. C. Saunders. 1987. Intraspecitic phylogeography: the rnitocho~ DNA bridge between population genetics and systematics. AM. Rev. Ekol. Syst. 18:489-522,
Baker, C. S., S. R. Palumbi, R. H. Lambemen, M. T. Weinrich, J. Calambokidis, and S. J. O'Brien. 1990. Muence of seasonal migration on geographic distribution of mitochondrial DNA haplotypes in humpback whales. Nature 344:238-240.
Baker, R. H. 1984. Origin, classification, and distn'bution. 4. 1-18 in L. K. Halls [ed.] White-tailed Deer Ecology and Management. Stackpole, Harrisburg, P~M.
Ballinger, S. W., L. H. Blankenship, J. W. Bickham, and S. M. Carr. 1992. AUoyme and mitochondria1 DNA analysis of a hybrid zone between white-tailed deer and mule deer (Odocoileus) in West Texas. Biochemical Genetics 30: 1-11.
Bartlett, S. E. and W. S. Davidson. 1991. Identification of Thwuurs tuna species by the polymerase chain reaction and direct seqwnce analysis of their mitochondria1 cytochrome b genes. Can. I. Fish. Aquat. Sfi. 48:3O9-3 17.
Boskovic, R., K. M. Kovacs, M. 0. Hammill, and B. N. White. 1996. Geographic distribution of mitochondrial DNA haplotypes in grey seals (Halichoerus grypus). Can. J, Zool. 74:1787-17%.
Brown, W. M., M. George, and A. C. Wilson. 1979. Rapid evolution of animal mitochondrial DNA. Proc. Nau. Acad. Sci. USA 76: 1967-1971. Bumell, F. L. and A. S. Harestad. 1983. Dispersal and dispersion of black-tailed deer: models and observations. J. Mammal. 64:201-209,
Cabot, E. L. and A. T. Beckenbach 1989. Simultaneous editing of multiple nucleic acid sequences with ESEE. Comput. Appl. Biosci. 5:233-234.
Cam, S. M. and H. D. Marshall. 1991. A direct approach to the measurement of DNA sequence variation in fish populations: applications of the polymerase chain reaction to studies of Atlantic cod (Gadus morhua). Journal of Fish Biology 39(supplement A): 101-107.
Cam, S. M. and G. A. Hughes. 19%. Direaion of introgressive hybridization between species of North American deer (Odocoileus) as iderred from mitochondrial cytochrome b seqyence. L Munm. 74:33 1-342.
Cam, S. M., S. W. Ballinget, J. N. Derr, L. H. Blankenship, and J. W. BicLham. 1986. Mitochondria1 DNA analysis of hybridization between sympatric white-tailed and mule deer in west Texas. Proc. Natl. Acad. Sci, USA 83:9576-9580.
Cam, S. M., A- J. SnelIen, K. A. Howse, and J. S. Wroblewski. 1995. Mitochondrial DNA sequence variation and genetic stock structure of Atlantic cod (Gadusnwrhua) from bay and offshore locations on the Newfoundland continental shelf. Molecular Ecology 4:79-88.
Chomczynski, P. and N. Sacchi. 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenoI-chloroformextraction. Analytical Biochemistry 162: 156-159.
Cothran, E. G., T. K. Chesser, M. H. Smith, and P. E. Johns. 1983. Influence of genetic variability and maternal factors on fetal growth in white-tailed deer. Evolution 37~282-291.
Cowan, I. McT. and P. A. Johnston. 1962. Blood serum protein variations at the species and subspecies level in deer of the genus Odocoileus. Syst. 2001. 11: 131- 138.
Cronin, M. A. 1991. Mitochondrial and nuclear genetic relationships of deer (Odocoileus spp.) in western North America. Can. I. Zool. 69: 1270-1279. Cronin, M. A. and V-C. Bleich. 1995. Mitochondrial DNA variationamong populations and subspecies of mule deer in California. Calif. Fish. & Game 8 1 :45- -- 54.
Cronin, M. A., E. Vyse, and D. Cameron. 1988. Genetic relationships between mule deer and white-Wed deer in Montana. I. Wildl. Manage. 52:320-328.
Cronin, M. A., S. C- Arnstrup, G. W. Garner, and E. R. Vyse. 1991. Znterspecific and intraspecific mitochondria1 DNA variation ia North American bears (Ursus). Can. J. Zooi. 69:2985-2992.
Derr, J- N. 1991. Genetic interactions between white-tailed and mule deer in the southwestern United States. I. Wildl. Manage. 55:228-237.
Derr, J. N., D. W. Hale, D. L. Ellsworth, and J. W. Bickham. 1991. Fertility of an F, male hybrid of white-tailed deer (0docoiIeu.s virginianccs) x mule deer (0.hemionus). I. Reprod. Fert. 93:lll-117.
Ellsworth, D. L., R L. Honeycutt, N. J. Silvy, J. W. Bickham, and W. D. Klimstra. 1994. Historical biogeography and contemporary patterns of mitochondria1 DNA variation in white-tailed deer from the southwestern United States, Evolution 48: 122-136.
Gavin, T. A. and B. May. 1988. Taxonomic status and genetic purity of Colurnbian white-tailed deer. J. WW.Manage. 52: 1-10.
Geist, V. 1981. Behaviour: adaptive strategies in mule deer. Pp. 157-223 in 0. C. Wallmo red.] Mule and Black-tailed Deer of North America. Univ. of Nebraska, I;incoln and London.
Gyllensten, U. B., D. Wharton, A. Josefsson, and A. C. Wilson. 1991. Paternal inheritance of mitochondria1 DNA in mice. Nature 352:255-257.
Hartl, D. L. and A. G. Clark. 1989. Principles of Population Genetics, Second Edition. Sinauer Associates, Inc., Sunderland, Massachusetts.
Hawkins, R. E. and W. D. IUimstta. 1970. A preliminary study of the social organization of white-tailed deer. I. Wildl. Manage. 34:4074l9. Hughes, G. A. 1990. A molecuIar genetic analysis of hybridization between two species of deer (Odocoileus) in western Canada. M.Sc. &+is, Department of Biology, Memorial University of Newfoundiand, St. John's.
Hughes, G. A. and S. M. Carr. 1993. ReciprocaI hybridization between white-tailed deer (Odocoilm virginimurs) and mule deer (0. hemionus) h western Canada: evidence fkom serum albumin and mtDNA sequences. Can. I. Zool. 71524-530.
Irwin, D. M., T. D. Kocher, and A. C. Wilson. 1991. Evolution of the cytochrome & gene of mammals. 1. Mol. Evol- 32:128-144.
Jukes, T. H. And C. R. Cantor. 1969. Evolution of protein molecules. Pp. 21-132 in H. N. Muow.] Mammalian Protein Metabolism. Academic Press, New York.
Kennedy, P. K., M. L. Kennedy, and M. L. Beck. 1987. Genetic variability in white- tailed deer (Odixoileus virginianus) and its relationship to environmental parameters and herd origin (Cervidae). Genetica 74: 189-201.
Kimura, M. 1980. A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16: 111-120.
Kocher, T. Do, W. K. Thomas, A. Meyer, S. V. Edwards, S. PiUibo, F. X. ViUablanca, and A. C. Wilson. 1989. Dynamics of mitochondria1 DNA evolution in animals: amplification and sequencing with conserved primers. Proc. Natl. Acad. Sci. USA 86:6196-6200.
Kramer, A. 1973. Interspecies behavior and dispersion of two sympatric deer species. J. Wildl. Manage. 37:288-300.
Kumar, S., M. Tamura, M. Nei. 1993. MEGA: Molecular Evolutionary Genetics Analysis, version 1-0 1. Pennsylvania State University, College Park.
Kurt&, B. and E. Anderson. 1980. Pleistocene Mammals of North America. Columbia University, New York.
Lehman, N. and R. K. Wayne. 1991. Analysis of coyote mitochondria1 DNA genotype fkequencies: estimation of the effective number of alleles. Genetics 128:405416. xhman, N., A. Eisenbawer, K. Hansen, D. L. Mech. R. 0. Peterson, J- P. Gogan. and R K. Wayne. 1991. Ii&ogression of coyote mitochoodrial DNA into sympatric North American gray wolf populations. Evoiution 45: 104-1 19.
Li, W. -H. 1997. Molecular Evolution. Sinauer Associates, Inc., Sunderland, Massachusetts.
Lingle, S. 1992. Escape gaits of white-tailed deer, mule deer and their hybrids: gaits observed and patterns of limb coordination. Behaviour 122: 153-18 1.
Lingle, S. 1993. Escape gaits of white-tailed deer, rude deer and their hybrids: body configuration, biomechanics, and ktion. Can. I. 2001. 71:708-724.
Marchinton, R. L. and D. H. Hirth. 1984. Behaviour. Pp. 129-168 in L. K. Halls [ed.] Whitetailed Deer Ecology and Management. Stackpole, Harrisburg, Penn.
Martin, R. A. and A. D. Bamosky. 1993. Quaternary mammals aad evoIutionary theory: introductory remarks and historical perspective. Pp. 1-12 in R. A. Martin and A. D. Bamosky [eds.] MorphoIogiiaI Change in Quaternary Mammals of North America. Cambridge University Press, Cambridge.
McClymont, R. A., M. Fenton, and J. R. Thompson. 1982. Identification of cervid tissues and hybridization by senun albumin. J. Wildl. Manage. 46540-544.
McCullough, D. R 1987. The theory and management of Oddleus populations. 4. 535-549 h C. M. Wemmer [ed.] Biology and Management of the Cervidae. Smithsonian Institution Press, Washington D. C.
McElroy, D., P. Mom, E. Bermingham, and I. KornfieId. 1991. REAP: The Restriction Enzyme Analysis Package, version 4.0. University of Maine, Orono.
Miyamoto, M. M., F. Kraus, P. L Laipis, S. M. Tanhauser, and S. D. Webb. 1993. Mitochoodrial DNA phylogenies within Artiodactyla. 4.268-281 in F. S. Szalay, M. J. Novacek, and M. C. McKenna [eds.] Mammal Phylogeny: Placentals. Springer-Verlag New York. hc., New York.
Moritz, C., T. E. Dowling, and W. M. Brown. 1987. Evolution of animal mitochondria1 DNA: relevance for population biology and systematics. Ann. Rev. Eco~.Syst. 18:269-292. Nei, M. 1985. Human evolution at the molecular level. 4.41-64 in T. Ohta and K. -. Ao ki [eds .] Population Genetics and Molecular Evolution. Japanese Scientific Societies Press, Tokoyo.
Nei, M. 1987. Phylogenetic trees. Pp. 287-326 in MoIecular Evolutionary Genetics. Columbia Press,
Nei, M. and F. Tajirna. 1981. DNA polymorphism detectable by restriction endonucieases. Genetics 97: 145-163.
Nelson, M. E. and L. D. Mech 1984- Home-range formation and dispersal of deer in northeastern Miisota, J- Mammal. 65567-575-
Nelson, M. E. and L. D. Mcch. 1992. Dispersal in female white-tailed deer. I. Mammal. 73 :891-894.
Palumbi, S. R. and C. S. Baker. 1994. Contrasting population structure from nuclear intron sequences and mtDNA of humpback wbales. Mol. Biol. Evol. 1l:426-435.
Pamilo, P. and M. Nei. 1988. Relationships between gene trees and species trees. Mol. Biol. Evol- 5568-583.
Pielou, E. C. 1991. After the Ice Age: The Return of Life to Glaciated North America. The University of Chicago Press, Chicago.
Ramsey, P. R., J. C. Avise, M. H. Smith, and D. F. Urbston. 1979. Biochemical variation and genetic heterogeneity in South Carolina deer populations. I. Wildl. Manage. 43:136-142.
Saitou, N. and M. Nei. 1987. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:4064ZS.
Sheffield, S. R., R. P. Morgan, G. A. Feldhamer, and D. M. Harman. 1985. Genetic variation in white-tailed deer (Odocoileus virginianus) popdations in western Maryland. I. Md.66:243-255.
Slade, R. W., C. Moritz, A. Heideman, and P. T. Hale. 1993. Rapid assessment of singlecopy nuclear DNA variation in diverse species. Molecular Ecology 2:359- 373. Smith. W. P. 1987. Dispersion and habitat use by sympaeic Columbian white-taiIed deer and Columbian black-tailed deer- J. Mammal. 68:337-347.
Strimmer, K. and A. von Haeseler. 1996. Quartet puzzling: a quartet maximum Likelihood method for reconstructing tree topologies. Mol. Biol. Evol. 13:964-969.
Stubblefield, S., R. Wamn, and B. Murphy. 1986. Hybridization of Bee-ranging white-tailed and mule deer in Texas. J. Wildl. Manage. 50:688690.
Swenson, I. E., S. I. Knapp, and H. J. Wentland. 1983. Wirdistribution and habitat use by mule deer and whitetailed deer in southeastern Montana. Prairie Nat. 15:97-112.
Swofford, D . L. 1996. PAUP: Phy logenetic Analysis Using Parsimony. Version 4.0.0d51 and subsequent versions.
Swofford, D. L. and D. P. Begle. 1993. PAUP: Phylogenetic Analysis Using Parsimony. Version 3.1 User's Manual. Srnithsonian Institution, Washington, DC.
Swofford, D. L., G. J. Olsen, P. J. Waddell, and D. M. Hillis. 1996. Phylogenetic inference. Pp. 407414 in D. M. Hitlis, C. Moritz, and B. K. Mable [eds.] Molecular Systematics, Second Edition. Sinauer Associates, k.,Sunderland, Massachusetts.
Tarnura, K. and M. Nei. 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondria1 DNA in humans and chimpanzees. Mol. Biol. Evo~.105 12-526.
Tierson, W. C., G. F. Maafeld, R W. Sage, and D. F. Behrend. 1985. Seasonal movements and home ranges of white-tailed deer in the Adirorsdacks. J. Wildl. Manage. 49:760-769.
Wallmo, 0.C. 1981. Mule and black-tailed deer distribution and habitats. Pp. 1-25 in 0.C. Wallmo Ed.] Mule and Black-tailed Deer of North America. University of Nebraska, Lincoln and London-
Wenink, P. W., A. I. Baker, HI-U. Roswr, and M, G. J. Tilanus. 1996. Global mitochondrial DNA phylogeography of holarctic breeding dunlins (Calidris abimz). Evolution 50:3 18-330. Whitehead, G. K. 1972. Deer of the World. Constable and Company Ltd., London.
Wilson, A. C.,R. L. Cam, S. M. Cam. M-George, U. B. Gyllensten, K. M. Helm- Bychowski, R. G. Higuchi, S. R. Palumbi, E. M. Prager, R. D. Sage, and M. Stoneking. 1985. Mitochondria1 DNA and two perspectives on evolutionary genetics. Biol. J. Linn. SIC.26:375-400-
Wilson, D. E. and D. M. Reeder [eds.] 1993. Mammal Species ofthe World: A Taxonomic aod GeograpKc Reference, Second Edition. Smithsonian Institution Press, Washington D. C.
Wishart, W. D. 1980. Hybrids of white-tailed and medeer in AIberta. J. Mamm. 61:716-720-
Zhang, D. and G. M. Hewitt. 1996. Highly collse~ednuclear copies of the mitochondria1 control region in the desert locust Schistocerca gregan'a: some implications for population studies. MoIecutar Ecology 5:295-300.
Zink, R. M. and D. L. Dittmstllll. 1993. Gene flow, refugia, and evolution of geographic variation in the song sparrow (Melopi' mefodia)).Evolution 47:717- 729.